ACTION OF CUERENTS ON THE SHOKE. 253 



place conclude that subsidence alone has not here sufficed to 

 produce tlie forms of the northern reef of this archipelago ; 

 since, for instance, it must have caused the isolated blocks of 

 coral in the southern channel of Kossol to form high reefs, just 

 as much and in the same way as has happened in Kriangle or 

 on Kossol itself. 



Thus we are obliged, under all the circumstances, to assume 

 the co-operation of some other force besides subsidence when en- 

 deavouring to explain the peculiar formation of the northern 

 reef, but still, without wholly excluding the effects of subsidence. 

 That force, as I believe, can only be sought in the action of the 

 constant currents of the sea ; for no other influence which might 

 check the growth of coral — such as sand, ooze in the water, 

 strong streams of freshwater, &c. — can be adduced in explanation 

 of the conditions and behaviour of this reef, nor have they ever 

 come into play in these spots. But the currents which during 

 rising tides here run strongly to the east, and at ebbing tides to 

 the west, might very well have acted on the original furrows 

 between Kriangle, Kossol, and Babelthuap, so as to widen 

 them to channels as these first sank below the surface of the sea, 

 and so have opened the way to the influx of the tidal currents. 

 Thus, combined with the currents, subsidence might have pro- 

 duced the form exhibited by the northern reef. However, it 

 has always seemed to me to be a very weak point in the 

 theory of subsidence, that it is evidently insufficient to explain 

 some particular cases, and requires to be supplemented by some 

 auxiliary force standing in no direct causal connection with the 

 theory itself; and I am convinced that many similar instances 

 will be found as investigators begin to take the trouble to study 

 separate reefs more exactly than has hitherto been done. 



A second difficulty occurs as follows. I have already stated 

 that tho triangle lying between Pelelew, XJrudzapel, and the 

 Urulong Channel, constituting nearly a fifth of the super- 

 ficies of the whole group, is almost level, and lies, in section, at 

 from two to four fathoms under water at high tide. This level 

 bottom consists of solid limestone which, rising gradually, con- 

 stitutes the numerous islands, composed of the same stone, 

 which are scattered about the suiface. It is cut through by 



